What if teeth… was tongue? These are the advanced questions we can finally answer with modern AI technology.

AI Horde now has lora support, but I think it is still limited to whatever the workers have decided to make available, instead of letting you ship a lora with your request.

If Collab has stopped handing out free GPUs, you might need to either find someone else handing out free GPU time, or else set up for slower CPU rendering. It is apparently possible to install the A1111 UI locally and run all the inference on the CPU, and that would be free and not limited in terms of models/loras/plugins, at the cost of speed.

You can use https://github.com/tsaost/xpng or https://github.com/receyuki/stable-diffusion-prompt-reader or even a text editor to read the prompt data embedded in the PNG file.

It is in the image! Along with the other settings. The A1111 web UI or other compatible tools should pull it out.